A transformer-based semi-autoregressive framework for high-speed and accurate de novo peptide sequencing
peer-reviewed · Communications Biology · 2025
| Date | 2025-02-14 |
| Type | peer-reviewed |
| Venue | Communications Biology |
| Publisher | Springer Science and Business Media LLC |
| Contribution | algorithm |
| DOI | 10.1038/s42003-025-07584-0 |
| Citations (OpenAlex) | 9 |
| Venue 2-year citedness | 6.33 |
Abstract
De novo peptide sequencing directly identifies peptides from mass spectrometry data, playing a critical role in discovering novel proteins and analyzing complex biological samples without reliance on existing databases. To address challenges in both speed and accuracy, a transformer-based model, TSARseqNovo, incorporates two key innovations: a Semi-Autoregressive decoder for parallel prediction of multiple amino acids and a Masking Refinement decoder for refining low-confidence predictions. These features significantly enhance sequencing efficiency and accuracy. Evaluations on the Nine-Species, Aggregated, and Glycoproteomic datasets, demonstrate that TSARseqNovo outperforms state-of-the-art models, including CasaNovo, NovoB, InstaNovo + , and π-HelixNovo. Specifically, TSARseqNovo achieves up to a 2-fold speed increase over CasaNovo and π-HelixNovo, and approximately 10-fold over NovoB and InstaNovo + , while also showing substantial improvements in peptide prediction precision, especially for long peptides. These advancements position TSARseqNovo as a powerful tool for accelerating high-throughput proteomics research and addressing increasingly complex biological questions. A novel model for analyzing complex biological samples without reliance on databases has been proposed, demonstrating a 2- to 10-fold increase in speed and improved peptide identification precision compared to the current state-of-the-art model.
Methods and tools
- TSARseqNovo: Semi-autoregressive
Data used
- BALF proteomics - In-depth proteomic analysis of human bronchoalveolar lavage fluid towards the biomarker discovery for (as deposited) · PXD012645
- Casanovo data set and model weights (as deposited) · 10.5281/zenodo.6791263
- Diabetes causes marked inhibition of mitochondrial metabolism in pancreatic β-cells (as deposited) · PXD012979
- High-resolution spatially-resolved proteome mapping using automated, sacrificial liquid-mediated sample transfer from la (as deposited) · PXD008844
- Low-density lipoprotein receptor-related protein 1 (LRP1)-derived peptides protect against aggregation of LDL and choles (as deposited) · PXD011246
- Proteome of the rodent malaria parasite Plasmodium berghei liver stage merosomes (as deposited) · PXD010559
- Proteomic analysis of six different tissues from the Atlantic bottlenose dolphin (Tursiops truncatus) (as deposited) · PXD008808
- Simply extending the EThcD MS/MS range increases the confidence in N-glycopeptide identification. (as deposited) · MSV000083710
Cites (15)
- Bidirectional de novo peptide sequencing using a transformer model (2024) both
- Introducing π-HelixNovo for practical large-scale de novo peptide sequencing (2024) crossref
- Accurate de novo peptide sequencing using fully convolutional neural networks (2023) both
- De novo peptide sequencing with InstaNovo: Accurate, database-free peptide identification for large scale proteomics experiments (2023) both
- Introducing PandaNovo for practical large-scale de novo peptide sequencing (2023) semanticscholar
- De novo mass spectrometry peptide sequencing with a transformer model (2022) crossref
- Computationally instrument-resolution-independent de novo peptide sequencing for high-resolution devices (2021) both
- Personalized deep learning of individual immunopeptidomes to identify neoantigens for cancer vaccines (2020) both
- De novo sequencing of proteins by mass spectrometry (2020) both
- Peptide Sequencing with Deep Learning (2020) crossref
- DeepNovoV2: Better de novo peptide sequencing with deep learning (2019) semanticscholar
- De novo peptide sequencing by deep learning (2017) both
- Novor: Real-Time Peptide de Novo Sequencing Software (2015) both
- PEAKS: powerful software for peptide de novo sequencing by tandem mass spectrometry (2003) crossref
- Sequence database searches via de novo peptide sequencing by tandem mass spectrometry (1997) both
Cited by (4)
- Transformer Architectures for De Novo Peptide Sequencing and Peptide Property Prediction in LC–MS/MS Proteomics (2026) both
- Systematic benchmarking of mass spectrometry-based antibody sequencing reveals methodological biases (2025) both
- RT-GCTnovo: A Peptide De Novo Sequencing Model Incorporating Gated Multi-scale Features and Dynamic Mass Masks (2025) crossref
- Cumulating MS Signal enables polyclonal antibody analysis (2025) both